Constraint-Induced Movement therapy (CIMT) is one of few techniques that has been shown in controlled studies to produce large improvements in real-world arm function in stroke survivors with mild to moderate hemiparesis. The standard treatment involves one-on-one training of paretic arm use in the clinic for 3 hr/day for 10 consecutive weekdays and placing a mitt on the other hand to discourage its use. The considerable cost of CIMT, due to the demand on therapist time and requirement that patients travel daily to a clinic for 2 weeks, imposes substantial limits on those who might access the therapy. Our lab, with Peter Lum, Ph.D., has developed a device for automating the delivery of CIMT named AutoCITE; it permits practice of 8 upper- extremity tasks and provides frequent graphical performance feedback. Two reports indicate that AutoCITE is as efficacious as standard CIMT even with therapist effort reduced by 75%. A third report suggests that AutoCITE under simulated telerehabilitation conditions (i.e., patient and therapist were in different rooms in the lab) is also just as efficacious. We propose to modify AutoCITE so that it is more appropriate for use in patients' homes and evaluate its efficacy for providing CIMT on a telerehab basis. We will also specifically test whether making the AutoCITE software more interactive helps patients to comply with training at home, where there are more distractions. To this end, we will randomize 87 chronic stroke survivors to: 1. AutoCITE training in the home with two patients supervised concurrently by one therapist remotely via an Internet-based audio- visual link, 2. the same procedures except that the more interactive software will be run on AutoCITE, and 3. dose-matched, standard CIMT in the home. Motor outcomes and costs from the home-based, two-on-one, tele-AutoCITE groups will be compared to those from the home-based, one-on-one, in vivo CIMT group and to those from laboratory-based AutoCITE and standard CIMT groups run previously. If successful, AutoCITE, which involves simply automating task practice with actual objects and is much less expensive than prevailing virtual reality based approaches, might serve as a platform for delivering other upper-extremity therapies, in addition to CIMT, via telerehab. With about 300,000 new stroke survivors suffering motor disability annually, cutting the cost of and increasing access to efficacious upper-extremity rehabilitation has the potential to reduce disability significantly. PUBLIC HEALTH RELEVANCE Our laboratory, in collaboration with Peter Lum, Ph.D., has developed a device for automating the delivery of Constraint-Induced Movement therapy (CI therapy) named AutoCITE; it permits practice of 8 upper-extremity tasks and provides frequent graphical performance feedback. Two recently published studies suggest that AutoCITE is as efficacious as one-one-one CI therapy. Our proposed study will test the efficacy of the AutoCITE device to provide CI therapy on a tele-rehabilitation basis in patients' home. There are approximately 300,000 new strokes survivors with residual motor disability annually. Reducing the cost of and increasing access to efficacious upper-extremity rehabilitation has the potential to significantly reduce the burden of stroke on patients, their families and the healthcare system.